Collapsing SNP genotypes in case-control genome-wide association studies increases the type I error rate and power.

Collapsing SNP genotypes in case-control genome-wide association studies increases the type I error rate and power.